home / mcp / tmf620 mcp server

TMF620 MCP Server

Provides MCP-based access to TMF620 data for catalogs, offerings, and specifications with health monitoring.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "oopsyz-mcp": {
      "command": "uv",
      "args": [
        "run",
        "tmf620-mcp-server"
      ]
    }
  }
}

You run a modern MCP (Model Context Protocol) server that connects a mock TMF620 API with an AI agent environment. It exposes tools to explore and manage TMF620 data, enabling AI agents to query catalogs, product offerings, and specifications, and to monitor system health. This guide shows practical steps to install, configure, and use the MCP server end to end.

How to use

You interact with the MCP server through an MCP client or AI agent. Start the MCP server alongside the mock TMF620 API, then load the MCP configuration so your agent can discover and invoke tools like listing catalogs, retrieving a catalog, listing product offerings, and creating product specifications. Use the health tool to verify that the MCP server and the TMF620 API are reachable.

Typical usage patterns include starting both services, ensuring the MCP server can reach the TMF620 API, and then issuing tool calls from your agent such as list_catalogs, get_catalog, list_product_offerings, get_product_offering, create_product_offering, list_product_specifications, get_product_specification, create_product_specification, and health. The MCP layer will route your agent’s requests to the appropriate underlying API endpoints and return structured results.

How to install

Prerequisites you need before installing: - Python 3.11+ for the MCP server and the mock TMF620 API - uv for fast dependency management and runtime - Optional: Docker if you prefer containerized deployment

# macOS/Linux
curl -LsSf https://astral.sh/uv/install.sh | sh

# Windows (PowerShell)
powershell -c "irm https://astral.sh/uv/install.ps1 | iex"

# Or install via Python package for environments without the installer
pip install uv

Option A: Using uv (recommended). Start both services from separate terminals: - uv run tmf620-mock-server - uv run tmf620-mcp-server

Option B: Background processes. Start each service in the background from the same or separate terminals: - uv run tmf620-mock-server - uv run tmf620-mcp-server

Option C: Traditional Python. Install dependencies and run the Python scripts directly:
- pip install -r requirements.txt
- python mock_tmf620_api_fastapi.py
- python tmf620_mcp_server.py

Configuration and connections

Configure how the MCP server connects to the TMF620 API and how the MCP server itself runs. The following configuration examples assume the mock TMF620 API is available locally on the default ports.

{
  "mcp_server": {
    "host": "localhost",
    "port": 7701,
    "name": "TMF620 Product Catalog API"
  },
  "tmf620_api": {
    "url": "http://localhost:8801/tmf-api/productCatalogManagement/v4"
  }
}

{
  "server": {
    "host": "localhost",
    "port": 8801,
    "protocol": "http"
  },
  "features": {
    "enable_cors": true,
    "enable_docs": true,
    "enable_mcp": true
  }
}

Claude Desktop integration (MCP clients)

To connect a Claude Desktop client, configure the MCP server entries so the client can launch the MCP server process automatically.

{
  "mcpServers": {
    "tmf620-mcp": {
      "command": "uv",
      "args": ["run", "tmf620-mcp-server"],
      "cwd": "/path/to/tmf620-mcp-server"
    }
  }
}

Available MCP tools and basic usage

The MCP server exposes tools to manage the TMF620 catalog data. You can list catalogs, retrieve catalogs, list product offerings, create product offerings, and manage product specifications. There is also a health tool to verify the status of the MCP server and its API connection.

Troubleshooting and notes

If you encounter connectivity issues, verify that the TMF620 API is reachable at the configured URL and that the MCP server is listening on its port. Check logs for errors, and use the health endpoint to confirm internal status. Ensure the environment paths and working directories for the MCP client are correctly set when launching through Claude Desktop or other orchestration tools.

Available tools

list_catalogs

List all product catalogs available in the TMF620 dataset.

get_catalog

Retrieve a specific catalog by its ID.

list_product_offerings

List product offerings, with optional filtering by catalog ID.

get_product_offering

Get a specific product offering by its ID.

create_product_offering

Create a new product offering under a specific catalog.

list_product_specifications

List all product specifications.

get_product_specification

Get a product specification by ID.

create_product_specification

Create a new product specification.

health

Check the MCP server and TMF620 API health.